My predominant impression of the HDRS is that it could lend itself to misinterpretation and potential alienation of a client. As someone who has undergone trauma and tragedy and works with clients undergoing trauma and tragedy, the ordinal designations on the test beg qualification and scrutiny. An important caution with ordinal scales in general is interpreting and reporting them as interval scales (Hays, 2017, p. 101), and, in this case, the risk is pronounced with what seem to be arbitrary ordinal designations. For instance, “fidgetiness” and “playing with hands, hair, etc.” represents a one-point difference in measuring “agitation” (Hamilton, 1960). Even when interpreted alongside an interview, I could see where a high score on this test could intensify feelings of isolation for a client, leaving them feeling confused over how to answer questions, wondering whether they should have answered a certain question in a different way, and ultimately questioning their reality.
While there are factors that give me confidence in the inventory as a reliable and valid instrument, such as having multiple questions per symptom (Fernandez & Isenhart, 1998), the factors that limit my confidence in the inventory are its interpretation of symptoms. The inventory is known for challenges to construct validity such as containing symptoms that are not specific to depression (Fernandez & Isenhart, 1998). For instance, it links the severity of depression in part with the presence/absence of, severity of, and timing of insomnia. However, insomnia can be an indicator of a variety of other disorders including anxiety and mania, and I am not aware of specific features of insomnia being used to judge severity of other disorders. In fact, even Insomnia Disorder F51.01 does not have a severity specifier (APA, 2022).
My main question related to the instrument is how the interpretation of the data might be improved if Likert scales for each symptom were used instead of rank ordering clusters of symptoms under a single domain. Likert scales are common for ordinal data (Hays, 2017, p. 100), and I noticed the HAM-A uses Likert scale reporting. As someone potentially experiencing hopelessness and helplessness, being able to self-report the severity of my symptoms on a Likert scale would be more motivating than picking symptoms from a list that may not be exhaustive and I might perceive as having arbitrary numeric values.
I personally would not use this tool in my future work as a professional counselor. Fernandez & Isenhart (1998) state that it classifies the severity of depression rather than the type, and I do see how classifying severity of depression would be useful. However, I am accustomed to the FDI principles of judging severity, and I prefer those over the scales of measurement in this tool due to the limitations described above.